Software Engineer - AI Platform

Armory•New York, NY

139d•Onsite

About The Position

We are building an agentic mesh network for US healthcare, deploying multimodal systems that can autonomously handle non-clinical tasks and with seamless execution between entities. Our founding team is ex-Palantir with direct experience delivering mission-critical software to HHS, NIH, and other federal institutions. We are based in New York City and work in-office every day. We operate with an extremely flat structure. We ship, we learn, we compound. You will design and build core systems that power autonomous agent orchestration across the most critical vertical within our country. This is not a feature-team role. You will own real subsystems with the autonomy and accountability that most engineers don't see until years into their careers.

Requirements

2+ years building production software. You've shipped real systems — not just side projects, not demos.
Deep understanding of how to get compounding leverage from AI tools like Claude Code.
Excited by distributed agent systems — state machines, message guarantees, fault tolerance, consensus.
Drawn to pushing the envelope of Applied AI and putting the latest models into production.
Operate with extreme ownership. Identify the highest-leverage problem and solve it.
Drawn to a sense of mission and purpose.

Nice To Haves

You may not have designed distributed agent systems from scratch yet, but you've thought hard about them and want to.

Responsibilities

Design and build core systems that power autonomous agent orchestration.
Own real subsystems with autonomy and accountability.
Build agent runtime — execution engine, task scheduling, state management, exception handling, and human-escalation pathways.
Develop complex agent orchestration — multi-agent coordination, tool use, real-time decision-making pipelines that process clinical documents, voice interactions, and multimodal healthcare data.
Create federated interoperability protocols that enforce data residency while enabling cross-entity agent collaboration.
Develop developer-facing abstractions — APIs, SDKs, and workflow primitives that allow both internal and external teams to deploy agents on the platform.